How does "Chunk information before asking working memory to use it" work?

Working memory can hold roughly four to seven chunks, but a "chunk" can be arbitrarily large if it is already consolidated in long-term memory. Experts can perform complex operations that would overwhelm a novice because each move is a single practiced chunk, not a string of elements. Teaching or practicing component skills to automaticity before combining them grows effective working-memory capacity without exceeding its actual limit.

How does "Strip extraneous cognitive load from learning materials" work?

Cognitive load theory, which builds on Baddeley’s model, distinguishes intrinsic load (inherent task complexity), germane load (the effort of forming new schemas), and extraneous load (load caused by poor design, clutter, or distraction). Working memory is a fixed resource, so load spent on irrelevant processing directly reduces capacity for learning. Reducing extraneous load reallocates that capacity to actual schema formation.

How does "Integrate across modalities with narrative or story" work?

Baddeley later added the episodic buffer to his model: a limited-capacity store that integrates information from the phonological loop, the visuospatial sketchpad, and long-term memory into coherent episodes. Material that has narrative structure or contextual coherence occupies the buffer more efficiently than isolated facts, because the connections between items are stored once rather than independently.

How does "Offload peripheral information to free working memory" work?

Working memory is a processing system, not a storage system. Using it to hold information you do not actively need — a phone number while writing an email, the steps already done in a recipe, a list of errands — leaves less capacity for the task at hand. Offloading to paper or a trusted system eliminates the storage demand and lets processing proceed at full capacity.

How does "Train task-switching to reduce cognitive overhead" work?

The central executive — the supervisory component of working memory — bears the cost of switching between tasks: retrieving the current context, suppressing the previous one, and reorienting attention. This "switch cost" is real and measurable. Practicing transitions between specific, predictable task types reduces the reconfiguration time and the errors that occur in the switching window.

How does "Study worked examples before solving problems independently" work?

For novices, problem-solving without schema consumes working memory almost entirely on search strategies, leaving little for learning. Studying worked examples offloads the search process: the solution path is already given, so cognitive resources can be directed to understanding the structure. As schemas form, the worked-example advantage fades and problem-solving becomes appropriate — this is the "expertise reversal" transition.

Alan Baddeley

What is the working memory model and why does it matter for learning?

The phonological loop, visuospatial sketchpad, central executive, and what they mean for you

Short answer

Alan Baddeley and Graham Hitch’s working memory model describes the cognitive workspace that temporarily holds and manipulates information — it is not just a buffer but an active system with limited capacity. Understanding its architecture explains why multitasking degrades performance, why certain study designs backfire, and how to structure learning to fit the brain.

Working memory is the bottleneck through which all learning passes. Baddeley and Hitch proposed a model with distinct subsystems — a verbal loop, a visual sketchpad, an episodic buffer for integration, and a central executive that coordinates them. The model has been refined over decades and is supported by neuroimaging, lesion studies, and behavioral experiments. Understanding it is not academic: every practice below translates directly into fewer cognitive traffic jams when you are learning or thinking hard.

The practices (7)

Why it works

Baddeley’s model assigns spoken/written language to the phonological loop and visual-spatial material to the visuospatial sketchpad — two separate subsystems. Flooding both at once with redundant or competing content splits the central executive’s coordination resources and degrades processing. Using text for some information and images or speech for other information reduces interference between the two systems.

How to do it

1When designing materials, pair a narrated explanation with a diagram rather than placing the explanation as on-screen text.
2When studying, read one source at a time rather than toggling between multiple open tabs.
3Avoid playing spoken content (podcasts, videos) while reading dense text.

Evidence

Observational

The modality effect — that people learn better from narration plus images than from text plus images — is consistent with the dual-channel architecture of working memory and has been replicated across many multimedia learning studies.

Honest caveat: Effect sizes vary by material and by the learner’s prior knowledge; advanced learners often benefit less from segregated presentation.

— Mayer & Moreno (2003), "Nine ways to reduce cognitive load in multimedia learning", Educational Psychologist

Common mistake: Playing a lecture while also trying to read slides verbatim — the dual verbal input forces both into the phonological loop, splitting attention and reducing comprehension of both.

Go deeper on this practice →

Practice this with IX Coach

Reading about a practice changes nothing on its own. IX Coach turns these into a guided, adaptive routine — discerning where you are in real time and walking the practice with you, session after session.

Start with IX Coach →